OMSC and UMSC isolated from the human olfactory mucosa and human umbilical cord appeared as spindle-shaped cells and presented adipogenic and osteogenic differentiation potential as determined by Oil Red O and Alizarin Red S staining, respectively (Figure 1A). Phenotype analysis by flow cytometry showed that OMSC expressed CD29, CD44, CD73, CD90, CD105, and CD166 but did not express CD34 and CD45 (Figure 1B).

To check the therapeutic effect of OMSC and UMSC on EAE mice, their neurological function was evaluated daily, as shown in Figure 2A. During 31 days of EAE process, the neurological function of the mice treated with OMSC was significantly improved (p _OMSC-UMSC = 0.001,p _OMSC-PBS<0.001, Figure 2B). On days 16 to 20, daily clinical score differences among the OMSC, UMSC, and PBS treatment groups were statistically significant, showing the extraordinary therapeutic effects of OMSC. The day of EAE onset in each group was statistically significant (Figure 2C, F = 5.942, p = 0.015). The onset of the OMSC treatment group was later than that of UMSC treatment and control groups (Figure 2C; Table 1, p _OMSC-PBS = 0.005, p _OMSC-PBS = 0.024). The incidence within 30 days in OMSC treatment, UMSC treatment, and control groups were 66.7, 100, and 100%, respectively (Table 1), and the mortality were 16.67, 16.67, and 42.86% (Table 1), respectively. In conclusion, EAE mice had improved neurological function, delayed onset time, and reduced incidence and mortality rate after OMSC intervention.

Spinal cord sections of mice in OMSC, UMSC, and PBS treatment groups were stained with HE and LFB (Figure 2D), respectively. It was found that compared with PBS and UMSC groups, the inflammatory and demyelinating areas of the OMSC group showed a downward trend. However, there was no statistical significance in the histological inflammation score and demyelination score among OMSC, UMSC treatment groups, and control group (data not shown).

In order to study the effect of OMSC and UMSC on serum inflammatory factors in EAE mice, serum IFN-γ, TNF-α, IL-6, and IL-10 levels were detected by CBA. Differences of serum IFN-γ levels between OMSC treatment, UMSC treatment, and control groups were statistically significant (Figure 3, p _IFN-γ = 0.004). The serum levels of IFN-γ in UMSC treatment and OMSC treatment groups were lower than those in the control group (Figure 3, p _UMSC-PBS = 0.002, p _OMSC-PBS = 0.003). Serum levels of TNF-α, IL-10, and IL-6 in the three groups were not statistically significant (Figure 3, p _IL-10 = 0.120, p _TNF-α = 0.085, p _IL6 = 0.646).

We have found that OMSC better improved neurological function and reduced the serum IFN-γ level in EAE mice compared with UMSC. Since IFN-γ is a crucial cytokine secreted by Th1 lymphocytes, and Th1 lymphocytes, characterized as CD3+CD4+IFN-γ+ cells, play a pivotal role in the development of EAE; we assumed that early preventive treatment of OMSC could downregulate the proportion of CD4+IFN-γ+ T cells. We co-cultured UMSC, OMSC, and human peripheral blood mononuclear cells (PBMCs) and stimulated them with MOG_35-55 for 48 h to observe the immunomodulatory effects of UMSC and OMSC under MOG_35-55 stimulation. Since serum TNF-α showed a decreasing trend in EAE mice receiving OMSC treatment, and TNF-α is also an important marker for EAE, CD4+IFN-γ+ T cell proportion and CD4+TNF-α+ T cell proportion were all detected by flow cytometry (Figure 4A). The proportion of CD4+IFN-γ+ T cells (Figure 4B) and CD4+TNF-α+ T cells (Figure 4C) significantly decreased in MOG_35-55-stimulated PBMC co-cultured with UMSC and OMSC compared to MOG_35-55-stimulated PBMC cultured alone. Meanwhile, OMSC exhibited a stronger inhibitory effect on CD4+TNF-α+ T cells than UMSC (Figure 4C).

For the in vivo study, the mice underwent OMSC preventive treatment, and the control group was sacrificed at the day of EAE onset, and splenic lymphocyte analysis was performed. The proportion of CD4+IFN-γ+ T cells (Th1 cells) in the PBS and OMSC treatment groups was analyzed (Figure 4D). The results showed that the percentages of CD4+IFN-γ+ T cells in the OMSC treatment and PBS groups were 7.61 ± 1.96 % and 28.97 ± 0.54 %, respectively, and the difference was statistically significant (Figure 4E, t = 10.51, p < 0.001). Interestingly, although EAE induced by MOG_35-55 is mainly CD4+ T cell–dependent, the proportions of CD8+T lymphocytes were also lower in the OMSC treatment group, indicating a multi-immunomodulatory potential of the OMSC. There was no significant difference in the proportions of CD4+T lymphocytes and CD8⁺ IFN-γ+ T lymphocytes between the two groups. The above mentioned studies highlighted the immunomodulatory effects of OMSCs.

In order to study whether OMSC can regulate the T-cell subsets of human lymphocytes and their mechanism, we constructed a co-culture system of OMSC and human PBMC and detected the proportion of CD4+IFN-γ+ T cells by flow cytometry (Figure 5). The results showed that after 48 h of co-culture, the proportion of CD4+IFN-γ+ T cells in the single culture group (Figure 5A) was 6.69 ± 0.39 % and that in the co-culture group (Figure 5B) was 1.14 ± 0.24 %. There was a significant difference in the proportion of CD4+IFN-γ+ T cells between the two groups (Figure 5E, F = 119.60, p < 0.0001), which confirmed the immunomodulatory effects of OMSC.

Besides, we used indomethacin and TGF-β1 inhibitors to elucidate the possible mechanism of OMSC that inhibited human CD4+IFN-γ+ T lymphocytes. Flow cytometry (Figures 5A–E) was used to detect the proportion of CD4+IFN-γ+ T cells in different groups. The results showed that the percentage of CD4+IFN-γ+ T cells in the COX inhibitor group (indomethacin) was 2.21 ± 0.12 % (Figure 5C) and that in the TGF-β1 inhibitor group was 1.25 ± 0.07 % (Figure 5D), respectively. The ratio of CD4+IFN-γ+ T cells in the single culture group was significantly higher than that in the OMSC co-culture group (Figure 5E, p < 0.001); the ratio of CD4+IFN-γ+ T cells in the COX inhibitor group was higher than that in the OMSC co-culture group (Figure 5E, p = 0.014). TGF-β1 inhibition has no obvious effect on OMSC-mediated inhibition of CD4+IFN-γ+ T cells. Taken together, these results suggest that OMSCs decreased the proportion of CD4+IFN-γ+ T cells in vitro partially via COX activity.

The olfactory mucosa of healthy donors was obtained through a nasal endoscope under local infiltration anesthesia with 2% lidocaine. The olfactory mucosa was cut with sterile instruments, digested with collagenase IV at 37°C for 1 h, and then passed through a 70-μm cell sieve (BD, CA, United States). After centrifugation at 4°C, 300 g, the cell suspension was resuspended and transferred into a culture bottle. The low-glucose DMEM (Gibco, Grand Island, NY, United States) was prepared with 10% FBS (Gibco, Grand Island, NY, United States), 100 U/ml penicillin (Gibco, Grand Island, NY, United States), and 0.1 g/L streptomycin (Gibco, Grand Island, NY, United States) (Ge et al., 2016). At the fifth passage, surface molecular identification and multi-differentiation staining and cell suspension preparation were performed.

The umbilical cord tissue was taken from puerpera and washed with normal saline three times, with blood vessels removed. The umbilical cord tissue was cut and soaked in a solution containing 100 U/ml penicillin (Gibco, Grand Island, NY, United States) and 0.1 g/L streptomycin (Gibco, Grand Island, NY, United States). After digesting with collagenase I and hyaluronidase at 37°C for 5 h, centrifuging at 4°C, 300 g, and passing through a 70-μm cell sieve, the cells were resuspended and transferred into a culture bottle (Corning, NY, United States). The low-glucose DMEM (Gibco, Grand Island, NY, United States) was prepared with 10% FBS (Gibco, Grand Island, NY, United States), 100 U/ml penicillin (Gibco, Grand Island, NY, United States), and 0.1 g/L streptomycin (Gibco, Grand Island, NY, United States). At the fifth passage, surface molecular identification and multi-differentiation staining and cell suspension preparation were performed.

After peripheral blood was acquired, polysucrose solution (Serumwerk Bernburg AG, Alere Technologies, Oslo, Norway) and density gradient centrifugation were used to separate peripheral blood lymphocytes. After purification using red blood cell lysis buffer (Solarbio, Beijing, China), the peripheral blood lymphocytes were resuspended and counted, and cultured in a CO₂ incubator at 37°C.

OMSC (or UMSC) were transferred into a 24-well plate (Thermo Fisher Scientific, Waltham, MA, United States) at a density of 1×10⁵ cells/well; CD3+ lymphocytes were purified using CD3 MicroBeads (Catalog # 130-050-101, Miltenyi Biotec, Bergisch Gladbach, Germany) and transferred into a 24-well plate at a density of 5×10⁵ cells/well. The total culture volume was 500 μl. For the in vitro co-culture study, each well was stimulated with 12.5 μg MOG_35-55. For the in vitro mechanism study, each well was stimulated with 500 ng anti-CD28 (BD, CA, United States) and 100 ng anti-CD3 (BD, CA, United States). For the COX inhibitor group, indomethacin (1 uM, Sigma-Aldrich, St. Louis, MO, United States) was added to the culture system, and for the TGF-β inhibitor group, anti-TGF-β (1 ug/ml) antibody was added to the culture system. At the last 6 h, the cells were stimulated with PMA (50 ng/ml) and ionomycin (500 ng/ml). Brefeldin A (BFA, 10 μg/ml) was used to inhibit the secretion of cytokines. Each experiment was repeated three times (n = 3).

The cells were collected and washed with PBS (Gibco, Grand Island, NY, United States) for staining. The cells were resuspended with 100 μl staining buffer, stained with APC-CD8 (BD, CA, United States) at 4°C for 30 min, fixed with 4% paraformaldehyde (Boster Biological Technology, Wuhan, China), stained with PE-Cy7-IFN-γ (BD, CA, United States), permeabilized (Invitrogen, CA, United States) at 4°C for 30 min, and washed and resuspended with PBS.

The fifth passage of UMSC were digested with 0.25% trypsin (Gibco, Grand Island, NY, United States), and OMSC were digested with 0.125% trypsin at 37°C. The concentration of the cell suspension was 1×10⁶/ ml; it was filtered by a 70-μm sieve (BD, CA, United States) and transferred to flow tubes. UMSC and OMSC were labeled with PE-CD29 (BD, CA, United States), PE-CD90 (BD, CA, United States), PE-Cy7-CD34 (BD, CA, United States), PE-Cy7-CD45 (BD, CA, United States), APC-CD44 (BD, CA, United States), APC-CD166 (BD, CA, United States), FITC-CD73 (BD, CA, United States), and FITC-CD105 (BD, CA, United States) at room temperature for 30 min. The surface molecules were detected by flow cytometry.

The fifth passage of UMSC (or OMSC) were transferred into 6-well plates (Thermo Fisher Scientific, Waltham, MA, United States) at the density of 3×10⁵/well. After 80% density, the old culture medium was removed, with a culture medium for the induction of osteoblasts by a previous study (Lei et al., 2013). The induction culture was carried out in the incubator with 5% CO₂ under 37°C, and half of the medium was changed every 2 days. At the 21st day of culture, the medium was sucked out, and the cells were washed with PBS buffer three times, fixed with 4% paraformaldehyde (Boster Biological Technology, Wuhan, China) for 15 min, stained with 0.1% Alizarin Red S (Rolex-Bio, Guangzhou, China) for 10 min, and washed with PBS three times.

The fifth passage of UMSC (or OMSC) were transferred into 6-well plates at the density of 3×10⁵/well. After 80% density, the culture medium was removed, with a culture medium for the induction of lipoblasts by a previous study (Lei et al., 2013). After induction with medium A for 3 days, fat induction medium B was used for 1 day. After circulation for 21 days, the medium was sucked out, and the cells were washed gently with PBS once and fixed with 4% paraformaldehyde for 15 min. After staining the cells using 0.5% Oil Red O (Sigma-Aldrich, St. Louis, MO, United States) for 10 min, they were washed carefully with isopropanol (Guangzhou chemical reagent factory, Guangzhou, China), gently rinsed with PBS three times, and then observed and photographed under a microscope.

C57BL/6 female mice were provided by Charles River, Ltd. All animal experiments, breeding, and care were performed according to Animal Experiment Center Guidelines and approved by the Animal Ethics Committee of Sun-Yat sen University.

EAE was induced in 6 to 8-week-old female C57BL/6 mice. The mice were anesthetized using 10 g/L pentobarbital sodium (Sigma-Aldrich, St. Louis, MO, United States) at a dose of 50 mg/kg. The pedal reflex and tail pinching reflex were observed to ensure that the mice were anesthetized. The skin was prepared on the back of the mice, and the mice were labeled with ear markers. An emulsion containing 1 mg/ml myelin oligodendrocyte glycoprotein (MOG) _35–55 (Sigma-Aldrich, St. Louis, MO, United States) together with complete Freund adjuvant (CFA, Sigma-Aldrich, St. Louis, MO, United States) with 4 mg/ml of M. tuberculosis H37R (BD, CA, United States) was injected subcutaneously on both sides of the back of the mice. In addition, 500 ng of pertussis toxin (List Biological Laboratories, New Delhi, INDIA) was injected intraperitoneally on days 0 and 2. All mice were randomly allocated to the OMSC treatment (n = 6), UMSC treatment (n = 6), and PBS groups (n = 7).

The mice were scored daily for neurological function evaluation according to the 6-point EAE scale as mentioned in a previous study (Ben-Zwi et al., 2019): 0, asymptomatic; 1, partial loss of tail tonicity; 2, tail paralysis; 3, hind limb weakness; 4, hind limb paralysis; 5, 4-limb paralysis; and 6, death.

On the 9^th day after immunization, the third to fifth passage of OMSC and UMSC were digested, filtered, and counted. According to the counting results, an appropriate number of cells were resuspended to equal concentration (1×10⁶/200 μl) of the cell suspension. The mice were fixed on the tail vein injection instrument, and the injection points were treated with 75% alcohol (Guangzhou Chemical Reagent Factory, Guangzhou, China). A 1-ml insulin needle (BD, CA, United States) was used to inject OMSC, UMSC cell suspension (1×10⁶/mouse), or PBS. The mice in OMSC treatment, UMSC treatment, and control groups were injected with cells or PBS via the tail vein.

After the mice were sacrificed, PBS and 4% paraformaldehyde were used in turns to perfuse, and spinal cords were carefully removed and fixed with 4% paraformaldehyde for 24 h, and then dehydrated using 75–95% ethanol. And the tissue was cleared with xylene, and a paraffin-embedded tissue block was made and was cut into paraffin sections.

Hematoxylin and eosin (HE) staining: The paraffin sections were put into ethanol and xylene for gradient dehydration. The hematoxylin staining solution was added to dye for 5 min and then washed with running tap water, followed by eosin dye staining for 5 min. After dehydrating, the sections were sealed with a sealing agent. Inflammation was scored as follows: 0, no inflammatory cells; 1, a few scattered inflammatory cells; 2, organization of inflammatory infiltrates around blood vessels; and 3, extensive perivascular cuffing with extension into the adjacent parenchyma, or parenchymal infiltration without obvious cuffing.

Luxol fast blue (LFB) staining: The paraffin sections were put into ethanol and xylene for gradient dehydration. Myelin staining solutions A and B were preheated, and the slices were put into dye A, dyed for 3 h, and then taken out and washed with water. After being immersed in staining solution B, the differentiation was terminated. After dehydrating, the sections were sealed with a sealing agent. Demyelination was scored as follows: 0, none; 1, rare foci; 2, a few areas of demyelination; and 3, large (confluent) areas of demyelination (Calida et al., 2001).

After anesthetization, the mice were sacrificed and peripheral blood was collected through the inner canthus vein with a capillary collecting vessel. After centrifugation at 4°C for 10 min, the upper serum was transferred to a new EP tube. The standard sample was diluted, and the standard curve was made according to the CBA kit’s (BD, CA, United States) operation instructions. The concentration of IFN-γ, TNF-α, IL-17, IL-6, and IL-10 was detected by flow cytometry.

On the day of EAE onset, the mice were anesthetized and sacrificed. The spleen of mice of OMSC treatment and control groups was observed and ground (n = 3). Red blood cell lysis buffer was used to remove red blood cells. After washing and centrifugation, the lymphocytes were resuspended in 1640 medium (Gibco, Grand Island, NY, United States) and counted by using the cell counting plate. Purified lymphocytes were transferred into 24-well plates at a density of 1×10⁶ cells/well. The volume of the culture system was 500 μl. Five microliters of leucocyte activation cocktail (BD, CA, United States) was added to each well to stimulate the cells for 6 h. After washing with PBS twice, the cells were resuspended in 100 μl PBS. After staining with FITC-CD3 (BioLegend, San Diego, CA, United States) and APC-CD4 (BioLegend, San Diego, CA, United States) at 4°C for 30 min, the lymphocytes were fixed with 4% PFA at room temperature for 15 min, then permeabilized, and stained with PE-Cy7-TNF-α (BioLegend, San Diego, CA, United States) and PE-IFN-γ (BioLegend, San Diego, CA, United States) at 4°C for 30 min. After washing with PBS, the lymphocytes were resuspended with PBS. The proportions of T lymphocyte subsets were detected by flow cytometry. Each experiment was repeated three times (n = 3).

GraphPad Prism 8 and SPSS 20.0 were used for graphing and statistical analysis. All data were expressed as mean ± SEM. The Kruskal–Wallis H test was used to evaluate the neurological function score of mice among three groups; one-way ANOVA was used for detecting differences in the mean values of the three groups, and the LSD-t test was used for pairwise comparison. A p value < 0.05 was considered statistically significant.